Navigable maps of structural brain networks across species

Area of Science:

• Neuroscience
• Network Science
• Computational Biology

Background:

• Connectomes are spatially embedded networks shaped by evolution, physical constraints, and communication needs.
• Understanding the relationship between a connectome's structure and its spatial layout is crucial for neuroscience.
• Existing models often struggle to fully explain the structural organization of connectomes across diverse species.

Purpose of the Study:

• To investigate the relationship between the structure of connectomes and their spatial layout across different species.
• To evaluate the navigability of connectomes using decentralized navigation protocols in both Euclidean and hyperbolic spaces.
• To propose a new framework for representing and analyzing connectomes based on their effective geometry.

Main Methods:

• Utilized a decentralized navigation protocol with greedy routing (visiting nearest geometric neighbors).
• Measured the success rate and length of greedy paths compared to shortest paths in connectome topology.
• Analyzed connectome navigability in both Euclidean and hyperbolic geometric spaces.

Main Results:

• Found significant differences in Euclidean space navigability between mammalian and non-mammalian species.
• Demonstrated that hyperbolic space provides highly navigable maps for all species' connectomes.
• Showed exceptional congruence between hyperbolic distances and connectome structure across species.

Conclusions:

• Euclidean distances alone are insufficient to explain the structural organization of connectomes.
• Hyperbolic geometry offers a powerful and universal framework for representing connectomes, suggesting a new 'hyperbolic cartography' of the brain.
• Hyperbolic maps facilitate the study of decentralized communication processes in connectomes universally across species and scales.